[time-nuts] HP53132A vs SR625
magnus at rubidium.dyndns.org
Sun Mar 17 07:45:59 EDT 2013
On 03/17/2013 01:43 AM, Bob Camp wrote:
> The 53132 is indeed a fine counter. It's got another flaw though - right at 10 MHz the resolution takes a dive. If you are doing time nut stuff, that may be a significant issue.
The frequency averaging method can make use of the beating of the
reference clock against the incomming clock, to get more information on
the frequency difference. If the clocks are near each other, it takes
longer averaging to get the same effective resolution. You can expect
the same type of effect around frequency ratios like 3/2 etc.
This is expected as it is a systematic effect where most of the new
samples does not bring any new info as they pop in systematically in the
places where they don't give info. Adding noise (sloppy trigger) could
actually improve the precission, as it would combat the quantization effect.
More information about the time-nuts